Yield variability trends of winter wheat and spring barley grown during 1932–2019 in the Askov Long-term Experiment

Highlights

• Winter wheat showed lower yield variability than spring barley (1932–2005).

• Since 2006, wheat yield variability trend increased mainly under higher NPK rates.

• Spring barley also showed increasing, but less pronounced, yield variability trend.

• Cereal yields were more variable under higher than under reduced NPK rates.

• Cereal yields were more stable with animal manure than with NPK fertilization.

Abstract

Designing cropping systems with low yield variability (or high yield stability) is becoming increasingly important because of the ongoing climatic and agronomic challenges. The trends in yield variability under various agronomic managements can be evaluated by conducting long-term experiments. By using a novel statistical analysis method, we estimated the long-term yield variability trends for winter wheat and spring barley grown in the Askov Long-term Experiment (Denmark) with different rates of mineral fertilizers (½, 1, and 1½ NPK) and animal manure (½, 1, and 1½ AM). Yield data from 1932 to 2019 were analyzed using a mixed model approach with restricted maximum likelihood (REML)-based parameter estimates. Across all nutrient treatments, winter wheat showed lower temporal yield variability than spring barley in the first decade. However, particularly, since 2006, the wheat yield variability trend increased mainly under treatments with higher NPK rates (1 and 1½). Spring barley also showed an increasing trend in yield variability; however, compared to wheat, this trend was less pronounced. Therefore, wheat yields were less stable than barley yields during the last decade. Wheat and barley yields fluctuated more under higher NPK rates (1 and 1½) than under a reduced rate (½ NPK). In general, animal manure provided more stable yields than NPK for wheat as well as for barley. The long-term trends in yield variability of cropping systems with winter wheat and spring barley were affected by the choice of crop, nutrient source and application rate, and increasing climatic variability.

Introduction

Because of the concerns of climate change and global food security, the stability of agricultural systems has become as equally important as their productivity (Olesen et al., 2000). In recent years, analyses of yield stability have become more important since variations in climate are also associated with the changes in crop productivity (Lobell and Field, 2007; Najafi et al., 2018; Ray et al., 2015). Although global yield data show that productivity is stagnating in large parts of the world, trends in yield variability are more difficult to detect (Arata et al., 2020).

Analyzing trends of yield stability over long periods is crucial to capture the potential impacts of climate change and to elucidate the effects on crop management. In addition to crop choice, e.g., spring vs winter crops (Reckling et al., 2018), cropping system design and fertilization rates affect agroecosystem interactions that may play an important role in the stabilization of yields (Macholdt et al., 2019).

Resilient cropping systems can be designed by better understanding the potential trends in yield variability over time and their interactions with crop choice and management. This requires both field-level data to avoid aggregation biases (Popp et al., 2005) and novel statistical methods. Long-term agricultural field experiments (LTEs) have remarkable potential to contribute to a better understanding of yield stability (Macholdt et al., 2020b, 2020a; Reckling et al., 2018; St-Martin et al., 2017). LTEs are defined as more than 20-year-old large-scale field experiments that study crop production, nutrient cycling, and environmental impacts of agriculture (Rasmussen et al., 1998). Typically, plot treatments in LTEs remain unchanged over long time periods (Cochran, 1939), allowing the investigation of the cumulative effects and processes, requiring several years to become evident, and their separation from weather effects and climate trends. Although these properties of LTEs make them ideal for quantifying temporal variations, LTEs are only recently being used more extensively to assess yield stability (Ahrends et al., 2018). Several hundred experiments are available: 620 are listed in a global assessment by (Debreczeni and Körschens, 2003), and, in Germany alone, Grosse et al. (2020) identified a total of 205 LTEs. This resource could be exploited more effectively in the future.

Treatments within LTEs may need to be adjusted over time to avoid fossilization of general management and changes in research focus. This renders the analysis of long-term trends challenging and demands creativity in data handling and methods of yield stability analysis (Barnett et al., 1995; Barnett and Riley, 1994). Novel statistical approaches involving mixed models can be used to generate residual maximum likelihood estimates of variances and means related to trend. By using these approaches, researchers can use stability indices such as Shukla’s stability variance to calculate temporal yield variability trends over time. While such approaches have been implemented for official variety trial data (Hadasch et al., 2020), they have not been used for analyzing yield data from LTEs and to reveal the effects of nutrient treatments on the changes in yield variability over time.

Herein, we estimate long-term yield variability trends for winter wheat and spring barley (1932–2019) grown with different rates of mineral fertilizer and animal manure. By applying a novel methodological approach, we tested the following hypotheses:

H1

There is an increasing trend toward more fluctuating cereal yields for winter wheat and spring barley over time.

H2

The yield variability of both crops is higher, and the increasing temporal trend of yield variability over time is greater at higher than at lower nutrient input levels.

H3

Animal manure reduces the yield variability of both crops and results in a lower increasing temporal trend of yield variability compared to mineral fertilizers.

H4

The yield variability of winter wheat is lower and the increasing temporal trend of yield variability is smaller than those for spring barley.